ljwolf

ljwolf

ljwolf

import scipy.stats as st
import copy
import numpy as np
import pysal as ps

def _estimate_kdes(y,z):
    """
    estimate a kernel desnity for each group in z (classes) from y (data)
    """
    z_unique = np.unique(z).tolist()

ljwolf

ljwolf

def joyplot(data, ax=None, 
            flatten = .1, #rescale the height of each distribution to avoid overlap. If large, will flatten out each of the KDEs
            linecolor='k', 
            shadecolor='w',
            alpha=1,
            shade=True, 
            line_kws = None,
            kde_kws=None,
            fig_kws=None,
            shade_kws=None):

ljwolf

ljwolf

select 
       t1.geom_uid as geom_uid, 
       t2.geom_uid as neighbor_uid
from 
       geo6311 as t1, 
       geo6311 as t2 
where 
       st_touches(t1.the_geom, t2.the_geom);
       
       

ljwolf

ljwolf

ljwolf

import numpy as np
import scipy.stats as st
import warnings

def retrodesign(theta, se, alpha=.05 , df=1, n_sims = 10000):
    """
    Implements the "retrodesign" function from Gelman \& Carlin (2014).

    This computes the power, Sign error rate, and exaggeration ratio.